Exposing French Agronomic Resources as Linked Open Data

Aravind Venkatesan, Nordine El Hassouni, Florian Philippe, Cyril Pommier, Hadi Quesneville, Manuel Ruiz, Pierre LarmandeSWAT4LS 2015

TL;DR

This paper introduces AgroLD, a project aiming to integrate diverse French agronomic data into a unified, queryable knowledge base using Linked Open Data principles. By exposing these resources as RDF, the project facilitates cross-domain research and knowledge discovery in agriculture.

What Problem Does This Paper Address?

The field of agronomy, encompassing genetics, plant molecular biology, and agro-ecology, is generating an overwhelming amount of data due to advancements in high-throughput technologies. This data is often siloed across various regional bioinformatics portals and databases, making it difficult for researchers to access, integrate, and analyze it effectively. This fragmentation hinders the extraction of new knowledge and limits the potential for comprehensive understanding of complex biological systems. The paper addresses this challenge by proposing a solution to expose these heterogeneous French agronomic resources as Linked Open Data, thereby promoting interoperability and enabling more powerful data integration and analysis.

Methods

The core of the AgroLD project is the application of Linked Open Data (LOD) principles. This involves transforming existing domain-specific data into Resource Description Framework (RDF) format. Key techniques include:

  • Ontology Alignment: Utilizing and integrating domain-specific ontologies (like Gene Ontology, Plant Ontology) to provide a common vocabulary for describing the data.
  • Metadata Integration: Incorporating metadata from various regional bioinformatics portals.
  • RDF Knowledge Base Creation: Building knowledge bases that link these diverse data sources using RDF, enabling semantic querying.
  • Phased Development: The project is being developed in phases, starting with data from the South Green Bioinformatics platform (SG) as a representative example.

Datasets

The AgroLD project integrates data from several French agronomic databases, with Phase I focusing on resources from the South Green Bioinformatics platform (SG). These include:

  • TropGeneDB: Hosts genetic, molecular, and phenotypic information on tropical crop species.
  • Ory GenesDB: A repository for functional genomics data in rice.
  • Oryza Tag Line: Contains sequence information (Flanking Sequence Tags) for rice.
  • GreenPhylDB: Provides sequence homology information for plant kingdom members.

Additionally, data from public sources like Gene Ontology, Plant Ontology, UniProtKB, and Gramene are integrated. The current AgroLD knowledge base includes data for species such as Oryza species, Arabidopsis thaliana, Sorghum bicolor, Zea mays, and Triticum species.

Key Insights and Ideas

  • The Power of Semantic Web for Biology: The paper highlights how Linked Open Data can effectively bridge the gap between disparate biological datasets, transforming them into a connected and queryable knowledge graph.
  • Enabling Cross-Domain Research: By making agronomic data interoperable, AgroLD has the potential to foster research that spans across genetics, molecular biology, and agro-ecology, leading to more holistic insights.
  • Leveraging Existing Ontologies: The strategic use of established ontologies is crucial for providing a standardized framework, ensuring that data from different sources can be meaningfully integrated and understood.
  • A Phased Approach to Complexity: The project's phased development strategy allows for iterative progress and validation, making the ambitious goal of integrating vast amounts of data more manageable.

Key Results

The paper presents the initial results of the AgroLD project, focusing on the creation of an RDF knowledge base that integrates data from the South Green Bioinformatics platform. While specific quantitative metrics like query performance or accuracy are not detailed in this introductory paper, the key qualitative result is the successful exposure of French agronomic resources as Linked Open Data. This lays the groundwork for future use cases and more in-depth analysis. The AgroLD knowledge base currently covers data for a selection of important crop and model plant species.

Limitations and Future Work

The authors acknowledge that this is an ongoing project. Future work includes:

  • Expanding Coverage: Integrating data from other regional bioinformatics portals beyond the South Green platform.
  • Extending Species Coverage: Including information for a wider range of plant species.
  • Developing Use Cases: Demonstrating the practical utility of the AgroLD knowledge base through real-world research applications.
  • Further Ontology Refinement: Continuously improving and extending the ontologies used to ensure comprehensive data representation.

Summary updated

BibTeX
@inproceedings{venkatesan2015exposing,
  author = {Aravind Venkatesan and Nordine El Hassouni and Florian Philippe and Cyril Pommier and Hadi Quesneville and Manuel Ruiz and Pierre Larmande},
  title = {{Exposing French Agronomic Resources as Linked Open Data}},
  booktitle = {SWAT4LS},
  year = {2015},
}
Figure